Quick notes from the 29th April teleconference:
0. Appoint a note taker for this call.
1. Any corrections, clarifications in the notes from the last meeting, and acceptance of the same.
2. Agenda bashing
3. Discuss work item 1a “Develop guidelines for machine readable files that must accompany submissions” led by Pete and Cory.
4. Brief discussion of next steps.
The chat room at http://webconf.soaphub.org/conf/room/omg-smsc was used for the call
Chair: Jishnu, Note taker: Jishnu. Notes were taken down in the chat room
Agenda Item 1: The notes from the previous call were accepted as published at http://www.omgwiki.org/smsc/doku.php?id=15_april_2010
Agenda Item 2: The Agenda was accepted without any change.
Agenda Item 3: The paper discussed was titled "Machine Readable Artifacts for OMG Submissions"
Pete gave an overview of the contents of this draft
Discussion of the item “The files should not arbitrarily change between versions, for example through use of a new tool (or versions thereof). That includes changes to identifiers.”
Larry: pointed out this may prove to be onerous for Domain submitters
Using a different tool to generate XSDs would often cause massive syntactic changes that have relatively little semantic change
Pete: pointed out that this is a “should” best practice rather than a shall “mandatory” requirement
Sridhar: If we doing our standards 'properly' the tool used should not change the interchange format - only a revision of the underlying model would cause a change
Sridhar: Suggested best practice : Interchange format changes should mirror model changes (small changes in model affect interchange by a small amount)
Larry: Unfortunately XMI is not specified in a way that makes it possible to achieve the second item mentioned by Sridhar
Pete: There are both XMI and UML issues
Larry took on the action item raise an issue for this
Pete: two issues: 1 against UML to make all associations ordered, and 2 an issue on XMI to require XMI serialization to reflect the same ordering as in UML
Doug: Verified that this discussion is limited to RTFs and the issues regarding change do not really apply to Submission adoption and FTF outputs since it is only at the end of the FTF process that version 1.0 of a standard becomes available.
Manfred: There are limitations in using tools like subversion for managing model versions
Discussing the Manifest issue brought up by Pete, Andrew and Jishnu said the info in the manifest is necessary. We can decide how it is presented in the published document
On the issue of zip files, Larry, Manfred and Doug suggested that two separate zip files one for normatives and one for informatives with each file containing a manifest of its content
Tom: When the files like schemas are published they need to be published at the right location URL
Jishnu: Contents of zip files will typically not need individual document numbers.
Collectively: We can most likely do with a single zip file per process with the manifest for the zip file which will be contained in the zip file mentioning things like whether a particular file in the zip file needs to be archived separately using its own document number or needs to be published at a specific URL.
Agenda Item 4:
Too many people are unavailable next week to make it worthwhile to meet. So the next meeting will be the following week.
Pete will send Jishnu an updated version of the document that was discussed today to place on the Wiki
Jishnu will propose an agenda for the next call before the end of this week, to be finalized next week.
Larry will file issues regarding UML and XMI deficiencies in consultation with Pete.
Keeping with the plan mentioned at the last meeting we will possibly spend our time primarily on item 1c - tooling led by Bob and JD.
Next Call Info:
Date/Time: Thursday 13th May 2010 18:00 BST, 13:00 EDT, 10:00 PDT
Number: +1 605 475 4850
Access code: 422074#
In case of problems getting onto the call, please call Andrew on +44 7710 469624.